[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6MK5BAWK2G53XWS2Z94TV6H":3,"skill-same:01M6MK5BAWK2G53XWS2Z94TV6H":79},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":23,"publishedAt":44,"linkOnly":45,"markdownContent":46,"markdownBody":47,"frontMatter":48,"refName":50,"commitSha":51,"skillDirectory":52,"forksCount":53,"repositoryHtmlUrl":54,"repositoryHomepageUrl":55,"ownerHtmlUrl":56,"ownerBlogUrl":57,"ownerDescription":58,"updatedAt":59,"install":60,"resources":72},"01M6MK5BAWK2G53XWS2Z94TV6H","flow-develop",null,"Multi-AI implementation using available external providers (Double Diamond Develop phase). DO NOT use for simple code edits, reading\u002Freviewing code, built-in commands, or trivial single-file changes.","hashgraph-online\u002Fawesome-codex-plugins","hashgraph-online","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F160371962?v=4","01M6CNB6NRJVNNGN14TR01Q6D3","HOL (Hashgraph Online)","01M6JT1HSKKG9V68KQAQGBA7XP",1575,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fnyldn\u002Fclaude-octopus\u002Fskills\u002Fflow-develop\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000030","开发工程师","PROFESSION","适用职业",[24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43],"flow","flow develop","develop","single","phase","simple","multi","true","single-file","Providers","commands","code","implementation","changes","file","reviewing code","use","for","do","reading","2026-11-10 10:12:30",false,"---\nname: flow-develop\ndescription: \"Multi-AI implementation using available external providers (Double Diamond Develop phase). DO NOT use for simple code edits, reading\u002Freviewing code, built-in commands, or trivial single-file changes.\"\ndisable-model-invocation: true\n---\n\n> **Host: Codex CLI** — This skill was designed for Claude Code and adapted for Codex.\n> Cross-reference commands use installed skill names in Codex rather than `\u002Focto:*` slash commands.\n> Use the active Codex shell and subagent tools. Do not claim a provider, model, or host subagent is available until the current session exposes it.\n> For host tool equivalents, see `skills\u002Fblocks\u002Fcodex-host-adapter.md`.\n\n\n{{PREAMBLE}}\n\nLoad `skills\u002Fblocks\u002Fengineering-method-selection.md` from the installed plugin\nand apply only the methods relevant to this task. Preserve this entry point's\nexecution contract and output format. Read referenced skills as instructions;\ndo not invoke the current command recursively or add provider calls from a seat.\n\n## Portable feature boundary\n\nBefore implementation, select the existing feature and run its boundary adapter:\n\nUse the installation that loaded this skill for both execution and validation.\nIf `CLAUDE_PLUGIN_ROOT` is absent, replace every `OCTO_ROOT` fallback in this skill with\nthe absolute plugin root derived from this loaded skill's path. Do not infer it\nfrom the project directory or an existing stable link. The root helper refreshes\ninstalled copies while preserving another session's working root when called\nfrom an inferred development checkout.\n\n```bash\nOCTO_ROOT=\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\"\nbash \"$OCTO_ROOT\u002Fscripts\u002Fhelpers\u002Ffeature-workflow.sh\" boundary develop \"${OCTOPUS_FEATURE:-}\"\n```\n\nRead the bound project policy, relative source and digest. Ask the returned marker batch once through the available native host question tool. Keep skipped or noninteractive decisions open. Report the open count and defer only tasks whose current contract depends on a named unanswered decision. The runtime repeats these checks before actual task dispatch.\n\nTwo or more spec\u002Fplan\u002Ftasks artifacts trigger deterministic analysis automatically. A clean pass uses no extra model seat. Unresolved findings can trigger at most one independent bounded seat, shared across host and runtime for that artifact revision. Unavailable or failed semantic review warns and proceeds. Analysis reports exact source quotations and never rewrites artifacts.\n\nUse stable task IDs and metadata when present. The parent resolves paths, dependencies and current Git state at every wave. A hint cannot grant concurrency or permit overwriting user changes. Legacy inputs retain their existing planner and execution path.\n\n## Pre-Development: State Check\n\nBefore starting development:\n1. Read `.octo\u002FSTATE.md` to verify Define phase complete\n2. Update STATE.md:\n   - current_phase: 3\n   - phase_position: \"Development\"\n   - status: \"in_progress\"\n\n```bash\n# Verify Define phase is complete\nif [[ -f \".octo\u002FSTATE.md\" ]]; then\n  define_status=$(\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Focto-state.sh\" get_phase_status 2)\n  if [[ \"$define_status\" != \"complete\" ]]; then\n    echo \"⚠️ Warning: Define phase not marked complete. Consider running definition first.\"\n  fi\nfi\n\n# Update state for Development phase\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Focto-state.sh\" update_state \\\n  --phase 3 \\\n  --position \"Development\" \\\n  --status \"in_progress\"\n```\n\n\n## ⚠️ EXECUTION CONTRACT (MANDATORY - CANNOT SKIP)\n\nThis skill uses **ENFORCED execution mode**. You MUST follow this exact sequence.\n\n### STEP 1: Detect Work Context (MANDATORY)\n\nAnalyze the user's prompt and project to determine context:\n\n**Knowledge Context Indicators**:\n- Deliverable terms: \"PRD\", \"proposal\", \"presentation\", \"report\", \"strategy document\", \"business case\"\n- Business terms: \"market entry\", \"competitive analysis\", \"stakeholder\", \"executive summary\"\n\n**Dev Context Indicators**:\n- Technical terms: \"API\", \"endpoint\", \"function\", \"module\", \"service\", \"component\"\n- Action terms: \"implement\", \"code\", \"build\", \"create\", \"develop\" + technical noun\n\n**Also check**: Does project have `package.json`, `Cargo.toml`, etc.? (suggests Dev Context)\n\n**Capture context_type = \"Dev\" or \"Knowledge\"**\n\n#### Step 1b: Detect Dev Subtype (if Dev context)\n\nWhen context_type is Dev, determine the **subtype** to inject domain-appropriate quality guidance into the prompt sent to providers. Append the matching supplement text after the user's prompt.\n\n| Subtype | Trigger keywords | Quality supplement |\n|---------|-----------------|-------------------|\n| `frontend-ui` | \"page\", \"widget\", \"component\", \"UI\", \"HTML\", \"CSS\", \"form\", \"dashboard\", \"layout\" | See **frontend-ui enrichment** below. |\n| `cli-tool` | \"CLI\", \"command-line\", \"terminal\", \"script\", \"flag\", \"argument\" | Help text via --help flag. Meaningful exit codes (0 success, 1 user error, 2 system error). Stdin\u002Fstdout\u002Fstderr used correctly. Argument validation with clear error messages. |\n| `api-service` | \"API\", \"endpoint\", \"REST\", \"GraphQL\", \"gRPC\", \"server\", \"route\" | Input validation at boundaries. Consistent error response format. Auth\u002Fauthz on every endpoint. Rate limiting consideration. OpenAPI\u002Fschema documentation. |\n| `infra` | \"deploy\", \"terraform\", \"docker\", \"CI\", \"pipeline\", \"Kubernetes\", \"helm\" | Idempotent operations. Secrets never hardcoded. Rollback path documented. Health checks included. |\n| `data` | \"ETL\", \"pipeline\", \"migration\", \"schema\", \"database\", \"SQL\" | Idempotent migrations. Backup\u002Frollback strategy. Data validation at ingestion. |\n| `general` | Default if no subtype matches | No supplement — use base implementer persona only. |\n\n#### frontend-ui enrichment\n\nWhen `frontend-ui` subtype is detected, do TWO things:\n\n**A. Inject quality supplement into the prompt:**\nSelf-contained files preferred. Accessibility: ARIA labels, keyboard nav, 44px touch targets (WCAG 2.5.5). Safe DOM: createElement over innerHTML. Progressive enhancement: feature-detect APIs (navigator.share, localStorage) with fallbacks. Persist user prefs via localStorage.\n\n**B. Pull design intelligence from BM25 (if available):**\n\nBefore calling orchestrate.sh, check if the design intelligence engine exists and query it for relevant design context:\n\n```bash\nSEARCH_PY=\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fvendors\u002Fui-ux-pro-max-skill\u002Fsrc\u002Fui-ux-pro-max\u002Fscripts\u002Fsearch.py\"\nif [[ -f \"$SEARCH_PY\" ]]; then\n    # Detect relevant domains from the prompt\n    design_context=\"\"\n    # Style query — what visual style fits this task?\n    style_hit=$(python3 \"$SEARCH_PY\" \"\u003Cuser's task description>\" --domain style --top 1 2>\u002Fdev\u002Fnull || true)\n    [[ -n \"$style_hit\" ]] && design_context+=\"Design style suggestion: $style_hit\\n\"\n    # UX query — relevant UX patterns\n    ux_hit=$(python3 \"$SEARCH_PY\" \"\u003Cuser's task description>\" --domain ux --top 1 2>\u002Fdev\u002Fnull || true)\n    [[ -n \"$ux_hit\" ]] && design_context+=\"UX pattern: $ux_hit\\n\"\n    # Append to prompt if hits found\n    if [[ -n \"$design_context\" ]]; then\n        # Append design intelligence to the orchestrate.sh prompt\n        prompt=\"${prompt}\\n\\nDesign intelligence (from BM25 search):\\n${design_context}\"\n    fi\nfi\n```\n\nThis gives providers concrete design guidance (style direction, UX patterns) without requiring the user to run `\u002Focto:design-ui-ux` separately. If the search engine isn't installed, implementation proceeds with the quality supplement only.\n\n**How to apply:** When calling orchestrate.sh in Step 4, append the quality supplement (and design intelligence if available) to the prompt:\n```\norchestrate.sh develop \"\u003Cuser prompt>\\n\\nQuality requirements for this deliverable:\\n\u003Csupplement text>\\n\u003Cdesign intelligence if found>\"\n```\n\n**DO NOT PROCEED TO STEP 2 until context determined.** Context type (Dev vs Knowledge) and dev subtype determine which quality supplements and design intelligence to inject — wrong context wastes provider credits on irrelevant analysis.\n\n\n### STEP 2: Display Visual Indicators (MANDATORY - BLOCKING)\n\n**MANDATORY: You MUST use the native shell command tool to run this provider check BEFORE displaying the banner. Do NOT skip it. Do NOT assume availability.**\n\n```bash\nbash \"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fhelpers\u002Fcheck-providers.sh\"\n```\n\n**Use the ACTUAL results below. PROHIBITED: Showing only \"🔵 Claude: Available ✓\" without listing all providers.**\n\nIf `OCTO_ALLOWED_PROVIDERS` is set, treat it as the source of truth for which providers may participate. Providers filtered out by that allowlist are intentionally reported as unavailable; do not invoke or recommend them in the workflow.\n\n\n**Display this banner BEFORE orchestrate.sh execution:**\n\n**For Dev Context:**\n```\n🐙 **CLAUDE OCTOPUS ACTIVATED** - Multi-provider implementation mode\n🛠️ [Dev] Develop Phase: [Brief description of what you're building]\n\nProvider Availability:\n🔴 Codex CLI: ${codex_status} - Code generation and patterns\n🟡 Antigravity CLI: ${agy_status} - Alternative approaches\n🧭 Antigravity CLI: ${agy_status} - Additional external-model challenge\n🔵 Claude: Available ✓ - Integration and quality gates\n\n💰 Estimated Cost: 0.02-0.10 USD\n⏱️  Estimated Time: 3-7 minutes\n```\n\n**For Knowledge Context:**\n```\n🐙 **CLAUDE OCTOPUS ACTIVATED** - Multi-provider implementation mode\n🛠️ [Knowledge] Develop Phase: [Brief description of deliverable]\n\nProvider Availability:\n🔴 Codex CLI: ${codex_status} - Structure and framework application\n🟡 Antigravity CLI: ${agy_status} - Content and narrative development\n🧭 Antigravity CLI: ${agy_status} - Additional external-model challenge\n🔵 Claude: Available ✓ - Integration and quality review\n\n💰 Estimated Cost: 0.02-0.10 USD\n⏱️  Estimated Time: 3-7 minutes\n```\n\n**DO NOT PROCEED TO STEP 3 until banner displayed.** The banner shows users which providers will run and what costs they'll incur — starting API calls without this visibility violates cost transparency.\n\n\n### STEP 3: Read Prior State (MANDATORY - State Management)\n\n**Before executing the workflow, read any prior context:**\n\n```bash\n# Initialize state if needed\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" init_state\n\n# Set current workflow\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" set_current_workflow \"flow-develop\" \"develop\"\n\n# Get prior decisions (critical for implementation)\nprior_decisions=$(\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" get_decisions \"all\")\n\n# Get context from discover and define phases\ndiscover_context=$(\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" get_context \"discover\")\ndefine_context=$(\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" get_context \"define\")\n\n# Display what you found (if any)\nif [[ \"$discover_context\" != \"null\" ]]; then\n  echo \"📋 Discovery phase findings:\"\n  echo \"  $discover_context\"\nfi\n\nif [[ \"$define_context\" != \"null\" ]]; then\n  echo \"📋 Definition phase scope:\"\n  echo \"  $define_context\"\nfi\n\nif [[ \"$prior_decisions\" != \"[]\" && \"$prior_decisions\" != \"null\" ]]; then\n  echo \"📋 Implementing with decisions:\"\n  echo \"$prior_decisions\" | jq -r '.[] | \"  - \\(.decision) (\\(.phase)): \\(.rationale)\"'\nfi\n```\n\n**This provides critical context for implementation:**\n- Technology stack and patterns decided\n- Scope and requirements defined\n- Research findings to inform implementation\n- If **claude-mem** is installed, its MCP tools (`search`, `timeline`, `get_observations`) are available — use them to check for related past implementation patterns\n\n**DO NOT PROCEED TO STEP 4 until state read.**\n\n\n### STEP 4: Execute orchestrate.sh develop (MANDATORY - Use Bash Tool)\n\n**You MUST execute this command via the native shell command tool:**\n\n```bash\n# Keep execution and report validation in the same shell invocation.\nOCTO_ROOT=\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\"\nbash \"$OCTO_ROOT\u002Fscripts\u002Fhelpers\u002Fensure-plugin-root.sh\" || exit 1\nRUN_MARKER=$(mktemp) || exit 1\nRUN_STATUS=0\nOCTOPUS_WORKFLOW_INVOCATION_ID=\"${RUN_MARKER##*\u002F}\" \"$OCTO_ROOT\u002Fscripts\u002Forchestrate.sh\" develop \"\u003Cuser's implementation request>\" || RUN_STATUS=$?\nREPORT_STATUS=0\nSYNTHESIS_FILE=$(\"$OCTO_ROOT\u002Fscripts\u002Fhelpers\u002Fworkflow-result.sh\" develop \"$RUN_MARKER\" \"$RUN_STATUS\") || REPORT_STATUS=$?\nrm -f \"$RUN_MARKER\"\n[[ \"$REPORT_STATUS\" == 0 ]] || exit \"$REPORT_STATUS\"\necho \"VALIDATION PASSED: $SYNTHESIS_FILE\"\ncat \"$SYNTHESIS_FILE\"\n```\n\n**CRITICAL: You are PROHIBITED from:**\n- ❌ Implementing directly without calling orchestrate.sh — single-model implementation misses alternative approaches and edge cases that external providers surface through independent analysis\n- ❌ Writing code without multi-provider perspectives\n- ❌ Claiming you're \"simulating\" the workflow\n- ❌ Proceeding to Step 4 without running this command\n\n**You MUST use the native shell command tool to invoke orchestrate.sh.**\n\n#### What Users See During Execution (v7.16.0+)\n\nIf running in Claude Code v2.1.16+, users will see **real-time progress indicators** in the task spinner:\n\n**Phase 1 - External Provider Execution (Parallel):**\n- 🔴 Generating code and patterns (Codex)...\n- 🟡 Exploring alternative approaches (Antigravity)...\n\n**Phase 2 - Synthesis (Sequential):**\n- 🔵 Integrating and applying quality gates...\n\nThese spinner verb updates happen automatically - orchestrate.sh calls `update_task_progress()` before each agent execution. Users see exactly which provider is working and what it's doing.\n\n**If NOT running in Claude Code v2.1.16+:** Progress indicators are silently skipped, no errors shown.\n\n\n### STEP 5: Verify Execution (MANDATORY - Validation Gate)\n\n**After orchestrate.sh completes, verify it succeeded:**\n\nStep 4 checks the workflow exit status and validates a fresh\n`tangle-validation-*.md` report in this host's current session results directory.\nIt rejects missing reports and failed or unknown completion states. The validator\nhonors `CLAUDE_PLUGIN_DATA` and `CLAUDE_OCTOPUS_WORKSPACE`.\n\nUse the validated path printed by Step 4. If a later tool call starts a new shell,\nset `SYNTHESIS_FILE` to that exact path. Do not search for a different report\nor rerun the workflow to recover a shell variable.\n\n**If validation fails:**\n1. Report error to user\n2. Show logs from `~\u002F.claude-octopus\u002Flogs\u002F`\n3. DO NOT proceed with presenting results\n4. DO NOT substitute with direct implementation — fallback to single-model implementation skips the multi-provider synthesis that catches design flaws early\n\n\n### STEP 6: Update State (MANDATORY - Post-Execution)\n\n**After synthesis is verified, record implementation details in state:**\n\n```bash\n# Extract key implementation decisions from synthesis\nimplementation_approach=$(head -50 \"$SYNTHESIS_FILE\" | grep -A 3 \"## Implementation\\|## Approach\" | tail -3 | tr '\\n' ' ')\n\n# Record implementation decisions\ndecision_made=$(echo \"$implementation_approach\" | grep -o \"implemented\\|using [A-Za-z0-9 ]*\\|chose to\\|pattern:\" | head -1)\n\nif [[ -n \"$decision_made\" ]]; then\n  \"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" write_decision \\\n    \"develop\" \\\n    \"$decision_made\" \\\n    \"Multi-AI implementation consensus\"\nfi\n\n# Update develop phase context\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" update_context \\\n  \"develop\" \\\n  \"$implementation_approach\"\n\n# Update metrics\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" update_metrics \"phases_completed\" \"1\"\n```\n\n**DO NOT PROCEED TO STEP 7 until state updated.**\n\n\n### STEP 7: Present Implementation Plan (Only After Steps 1-6 Complete)\n\nRead the synthesis file and present:\n- Recommended approach\n- Implementation steps\n- Code overview from all available perspectives\n- Quality gates results\n- Request user confirmation before implementing\n\n**After user confirms: Implement the solution using Write\u002FEdit tools**\n\n**Include attribution:**\n```\n*Multi-AI Implementation powered by Claude Octopus*\n*Providers: available external providers + 🔵 Claude*\n*Full implementation plan: $SYNTHESIS_FILE*\n```\n\n\n# Develop Workflow - Develop Phase 🛠️\n\n## ⚠️ MANDATORY: Context Detection & Visual Indicators\n\n**BEFORE executing ANY workflow actions, you MUST:**\n\n### Step 1: Detect Work Context\n\nAnalyze the user's prompt and project to determine context:\n\n**Knowledge Context Indicators** (in prompt):\n- Deliverable terms: \"PRD\", \"proposal\", \"presentation\", \"report\", \"strategy document\", \"business case\"\n- Business terms: \"market entry\", \"competitive analysis\", \"stakeholder\", \"executive summary\"\n\n**Dev Context Indicators** (in prompt):\n- Technical terms: \"API\", \"endpoint\", \"function\", \"module\", \"service\", \"component\"\n- Action terms: \"implement\", \"code\", \"build\", \"create\", \"develop\" + technical noun\n\n**Also check**: Does the project have `package.json`, `Cargo.toml`, etc.? (suggests Dev Context)\n\n**Step 1b: Detect Dev Subtype** — see EXECUTION CONTRACT Step 1b above for subtype table and quality supplements. Append the matching supplement to the prompt before calling orchestrate.sh.\n\n### Step 2: Output Context-Aware Banner\n\n**For Dev Context:**\n```\n🐙 **CLAUDE OCTOPUS ACTIVATED** - Multi-provider implementation mode\n🛠️ [Dev] Develop Phase: [Brief description of what you're building]\n📋 Session: ${CLAUDE_SESSION_ID}\n\nProviders:\n🔴 Codex CLI - Code generation and patterns\n🟡 Antigravity CLI - Alternative approaches\n🔵 Claude - Integration and quality gates\n```\n\n**For Knowledge Context:**\n```\n🐙 **CLAUDE OCTOPUS ACTIVATED** - Multi-provider implementation mode\n🛠️ [Knowledge] Develop Phase: [Brief description of deliverable]\n📋 Session: ${CLAUDE_SESSION_ID}\n\nProviders:\n🔴 Codex CLI - Structure and framework application\n🟡 Antigravity CLI - Content and narrative development\n🔵 Claude - Integration and quality review\n```\n\n{{VISUAL_INDICATORS}}\n\n\n**Part of Double Diamond: DEVELOP** (divergent thinking)\n\n```\n       DEVELOP (tangle)\n\n        \\         \u002F\n         \\   *   \u002F\n          \\ * * \u002F\n           \\   \u002F\n            \\ \u002F\n\n       Diverge with\n        solutions\n```\n\n## What This Workflow Does\n\nThe **develop** phase generates multiple implementation approaches using external CLI providers:\n\n1. **🔴 Codex CLI** - Implementation-focused, code generation, technical patterns\n2. **🟡 Antigravity CLI** - Alternative approaches, edge cases, best practices\n3. **🔵 Claude (You)** - Integration, refinement, and final implementation\n\nThis is the **divergent** phase for solutions - we explore different implementation paths before converging on the best approach.\n\n\n## When to Use Develop\n\nUse develop when you need:\n\n### Dev Context Examples\n- **Feature Implementation**: \"Build a user authentication system\"\n- **Code Generation**: \"Create an API endpoint for user registration\"\n- **Complex Builds**: \"Implement a caching layer with Redis\"\n- **Architecture Implementation**: \"Create a microservice for payment processing\"\n- **Integration Work**: \"Integrate Stripe payment processing\"\n\n### Knowledge Context Examples\n- **PRD Creation**: \"Build a PRD for the mobile onboarding feature\"\n- **Strategy Documents**: \"Create a market entry strategy for APAC\"\n- **Business Cases**: \"Build a business case for migrating to cloud\"\n- **Presentations**: \"Create an executive presentation on Q2 results\"\n- **Research Reports**: \"Build a competitive analysis report\"\n\n**Don't use develop for:**\n- Simple one-line code changes (use Edit tool)\n- Bug fixes (use debugging skills)\n- Code review tasks (use deliver-workflow or review skills)\n- Reading or exploring code (use Read tool)\n- Simple document edits (use Write tool)\n\n\n## Visual Indicators\n\nBefore execution, you'll see:\n\n```\n🐙 **CLAUDE OCTOPUS ACTIVATED** - Multi-provider implementation\n🛠️ Develop Phase: Building and developing solutions\n\nProviders:\n🔴 Codex CLI - Code generation and patterns\n🟡 Antigravity CLI - Alternative approaches\n🔵 Claude - Integration and refinement\n```\n\n\n## How It Works\n\n### Step 1: Invoke Tangle Phase\n\n```bash\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Forchestrate.sh\" develop \"\u003Cuser's implementation request>\"\n```\n\n### Step 2: Multi-Provider Implementation\n\nThe orchestrate.sh script will:\n1. Call **Codex CLI** with the implementation task\n2. Call **Antigravity CLI** with the implementation task\n3. You (Claude) contribute implementation analysis\n4. Synthesize approaches and recommend best path\n\n### Step 3: Review Quality Gates\n\nThe tangle phase includes automatic quality validation:\n- Code quality checks\n- Security scanning\n- Best practice validation\n- Implementation completeness\n\n### Step 3b: Handle Agent Continuation (v9.5 \u002F CC v2.1.77+)\n\nWhen orchestrate.sh output contains `AGENT_TEAMS_RESUME:${agent_id}:${task_id}:${role}:${phase}`:\n\n1. Send a message to the existing agent using `SendMessage({to: \"${agent_id}\", message: \"\u003Citeration prompt>\"})` — this auto-resumes a stopped agent automatically (CC v2.1.77+)\n2. Read the iteration prompt from the resume instruction JSON at `~\u002F.claude-octopus\u002Fagent-teams\u002F${task_id}.json`\n3. If SendMessage returns an error (agent GC'd or not found), fall back to creating a new agent with the full prompt from the instruction file\n4. After the agent responds, store its agentId via: `bridge_store_agent_id(task_id, agent_id)` in the bridge ledger\n\nNote: The `Agent(resume: ...)` parameter was removed in CC v2.1.77. Always use `SendMessage` for agent continuation.\n\n### Step 4: Read Results\n\nResults are saved to:\n```\n~\u002F.claude-octopus\u002Fresults\u002F${SESSION_ID}\u002Ftangle-validation-\u003Ctimestamp>.md\n```\n\n### Step 5: Implement Solution\n\nAfter reviewing all perspectives, implement the final solution using Write\u002FEdit tools.\n\n\n## Implementation Instructions\n\nWhen this skill is invoked, follow the EXECUTION CONTRACT above exactly. The contract includes:\n\n1. **Blocking Step 1**: Detect work context (Dev vs Knowledge)\n2. **Blocking Step 2**: Check providers, display visual indicators\n3. **Blocking Step 3**: Execute orchestrate.sh develop via native shell command tool\n4. **Blocking Step 4**: Verify synthesis file exists\n5. **Step 5**: Present implementation plan, get user confirmation\n6. **After confirmation**: Implement the solution using Write\u002FEdit tools\n\nEach step is **mandatory and blocking** - you cannot proceed to the next step until the current one completes successfully.\n\n### Task Management Integration\n\nCreate tasks to track execution progress:\n\n```javascript\n\u002F\u002F At start of skill execution\nTaskCreate({\n  subject: \"Execute develop workflow with multi-AI providers\",\n  description: \"Run orchestrate.sh develop for implementation\",\n  activeForm: \"Running multi-AI develop workflow\"\n})\n\n\u002F\u002F Mark in_progress when calling orchestrate.sh\nTaskUpdate({taskId: \"...\", status: \"in_progress\"})\n\n\u002F\u002F Mark completed ONLY after implementation finished\nTaskUpdate({taskId: \"...\", status: \"completed\"})\n```\n\n### Error Handling\n\nIf any step fails:\n- **Step 1 (Context)**: Default to Dev Context if ambiguous\n- **Step 2 (Providers)**: If all external providers are unavailable, suggest `\u002Focto:setup` and STOP\n- **Step 3 (orchestrate.sh)**: Show bash error, check logs, report to user\n- **Step 4 (Validation)**: If synthesis missing, show orchestrate.sh logs, DO NOT substitute with direct implementation\n\nNever fall back to direct implementation if orchestrate.sh execution fails. Report the failure and let the user decide how to proceed.\n\n### Implementation Plan Format\n\nAfter successful execution, present implementation plan with:\n   ```\n   # Implementation Plan: \u003Ctask>\n\n   ## Recommended Approach\n   [Your strategic recommendation based on all perspectives]\n\n   ## Implementation Steps\n   1. [Step-by-step implementation plan]\n   2. ...\n\n   ## Code Overview\n\n   ### Codex Approach\n   [Key implementation details from Codex]\n\n   ### Antigravity Approach\n   [Alternative considerations from Antigravity]\n\n   ### Final Implementation\n   [Your integrated solution]\n\n   ## Quality Gates\n   - [ ] Security validation: PASSED\u002FFAILED\n   - [ ] Best practices: PASSED\u002FFAILED\n   - [ ] Code quality: PASSED\u002FFAILED\n\n   Would you like me to proceed with implementation?\n   ```\n\n6. **Implement the solution**\n   After user confirms, use Write\u002FEdit tools to create the implementation:\n   - Create necessary files\n   - Write code based on synthesized approach\n   - Add tests if applicable\n   - Document the implementation\n\n\n## Example Usage\n\n### Example 1: Build Authentication System\n\n```\nUser: Build a user authentication system with JWT\n\nClaude:\n🐙 **CLAUDE OCTOPUS ACTIVATED** - Multi-provider implementation mode\n🛠️ Develop Phase: Building authentication system\n\n[Executes: \"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Forchestrate.sh\" develop \"Build a user authentication system with JWT\"]\n\n[After completion, reads synthesis and presents:]\n\n# Implementation Plan: User Authentication System\n\n## Recommended Approach\n\nBased on multi-provider analysis, I recommend a layered approach:\n1. Auth middleware for JWT validation\n2. Login\u002Fregister endpoints\n3. Token refresh mechanism\n4. Secure password hashing (bcrypt)\n\n## Implementation Steps\n\n1. Create `src\u002Fauth\u002Fmiddleware.ts` - JWT validation middleware\n2. Create `src\u002Fauth\u002Fcontroller.ts` - Login\u002Fregister endpoints\n3. Create `src\u002Fauth\u002Fservice.ts` - Business logic\n4. Create `src\u002Fauth\u002Futils.ts` - Token generation\u002Fvalidation\n5. Update `src\u002Froutes.ts` - Register auth routes\n\n## Code Overview\n\n### Codex Approach\n- Modern TypeScript with strict types\n- Express middleware pattern\n- Redis for token blacklisting\n- Comprehensive error handling\n\n### Antigravity Approach\n- Passport.js integration suggestion\n- Rate limiting on auth endpoints\n- Multi-factor auth consideration\n- Session management alternatives\n\n### Final Implementation\n- Hybrid: Modern TypeScript + Express patterns\n- JWT with refresh tokens\n- bcrypt password hashing\n- Rate limiting included\n- Optional MFA hooks for future\n\n## Quality Gates\n- ✅ Security validation: PASSED (bcrypt, secure tokens)\n- ✅ Best practices: PASSED (TypeScript, error handling)\n- ⚠️  Code quality: WARNING (consider adding request validation)\n\nWould you like me to proceed with implementation?\n\n[User: Yes, proceed]\n\n[Claude creates files and implements the solution]\n```\n\n### Example 2: Create API Endpoint\n\n```\nUser: Create an API endpoint for fetching user notifications\n\nClaude:\n🐙 **CLAUDE OCTOPUS ACTIVATED** - Multi-provider implementation mode\n🛠️ Develop Phase: Creating API endpoint\n\n[Executes tangle workflow]\n\n[Presents implementation plan with multi-provider perspectives]\n[Implements the endpoint after user confirmation]\n```\n\n\n## Quality Gates Integration\n\nThe tangle phase automatically runs quality checks via `.claude\u002Fhooks\u002Fquality-gate.sh`:\n\n```bash\n# Triggered after tangle execution (PostToolUse hook)\n.\u002Fhooks\u002Fquality-gate.sh\n```\n\n{{QUALITY_GATES}}\n\n\n## Integration with Other Workflows\n\nTangle is the **third phase** of the Double Diamond:\n\n```\nPROBE (Discover) → GRASP (Define) → TANGLE (Develop) → INK (Deliver)\n```\n\nAfter tangle completes, you may continue to:\n- **Ink**: Validate and deliver the implementation\n\nOr use standalone for implementation tasks.\n\n\n## Before Implementation Checklist\n\nBefore writing code, ensure:\n\n- [ ] All providers responded with implementation approaches\n- [ ] Quality gates evaluated (security, best practices, code quality)\n- [ ] User confirmed the implementation plan\n- [ ] File structure and architecture are clear\n- [ ] Dependencies identified and available\n- [ ] Tests planned (if applicable)\n\n\n## After Implementation: Auto Code Review & E2E Verification (MANDATORY)\n\n**After implementation completes and before presenting results to the user, you MUST launch two verification agents in parallel.** Do NOT skip this step or ask the user whether to run it — it is automatic.\n\n### Launch both agents simultaneously:\n\n**Agent 1 — Code Review (Sonnet):**\n```\nAgent(\n  model: \"sonnet\",\n  subagent_type: \"feature-dev:code-reviewer\",\n  background execution: true,\n  description: \"Code review: post-develop\",\n  prompt: \"Review the code changes from this development session. Focus on:\n1. Bugs, logic errors, security vulnerabilities\n2. Hidden dependencies or coupling issues\n3. Whether error handling covers failure modes\n4. Adherence to project conventions (check CLAUDE.md)\n\nCheck git diff for the changed files. Report only high-confidence issues.\"\n)\n```\n\n**Agent 2 — E2E Verification (Sonnet):**\n```\nAgent(\n  model: \"sonnet\",\n  background execution: true,\n  description: \"E2E test: post-develop\",\n  prompt: \"Run end-to-end verification of the development changes:\n1. Run the project's test suite (detect from package.json scripts, Makefile, or pyproject.toml)\n2. Verify no regressions in existing tests\n3. Check that new files are properly integrated (imported, registered, sourced)\n4. Verify the implementation matches the original task requirements\n\nReport: tests passed\u002Ffailed, any integration issues found.\"\n)\n```\n\n**After both agents complete:**\n- Present their findings to the user as part of the results\n- If the code reviewer found HIGH-confidence issues, flag them prominently\n- If tests failed, flag before the \"what next?\" prompt\n- Do NOT block on the review — present findings alongside results\n\nWHY: The user should never have to manually request a code review after development work. Fresh-eyes review from a different model (Sonnet vs Opus) catches issues the implementer is blind to. Running tests automatically catches regressions before the user discovers them.\n\n\n## After Implementation Checklist\n\nAfter writing code, ensure:\n\n- [ ] All files created\u002Fupdated\n- [ ] Code follows recommended patterns from synthesis\n- [ ] Code follows the project's existing commenting conventions (do not add comments unless asked or required by project style)\n- [ ] Security concerns addressed\n- [ ] Error handling implemented\n- [ ] Tests written (if applicable)\n- [ ] Lint\u002Ftypecheck commands run (detect from package.json, pyproject.toml, Cargo.toml, Makefile; if not found, ask the user)\n- [ ] If lint\u002Ftest commands were discovered and not documented in CLAUDE.md, suggest adding them\n- [ ] **Auto code review completed** (Sonnet agent)\n- [ ] **E2E verification completed** (Sonnet agent)\n- [ ] User notified of completion with review findings\n- [ ] Suggest running ink-workflow for validation\n\n\n## Cost Awareness\n\n**External API Usage:**\n- 🔴 Codex CLI uses your OPENAI_API_KEY (costs apply)\n- 🟡 Antigravity CLI uses your AGY_AUTH_TOKEN (costs apply)\n- 🔵 Claude analysis included with Claude Code\n\nTangle workflows typically cost 0.02-0.10 USD per task depending on complexity and code length.\n\n\n## Post-Development: Checkpoint\n\nAfter development completes:\n1. Run fresh targeted tests for the changed behavior through\n   `skill-verification-gate`; stop if they fail.\n2. Create checkpoint: `git tag octo-checkpoint-post-develop-$(date +%Y%m%d-%H%M%S)`\n3. Update `.octo\u002FSTATE.md` with completion and add a history entry with files modified.\n\n```bash\n# Enter this block only after fresh targeted tests pass.\ncheckpoint_tag=\"octo-checkpoint-post-develop-$(date +%Y%m%d-%H%M%S)\"\ngit tag \"$checkpoint_tag\" -m \"Post-develop checkpoint from embrace workflow\"\necho \"📌 Created checkpoint: $checkpoint_tag\"\n\n# Update state only after verification and checkpoint creation succeed.\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Focto-state.sh\" update_state \\\n  --status \"complete\" \\\n  --history \"Develop phase completed\"\n\n# Record files modified in this phase\nmodified_files=$(git diff --name-only HEAD~1 2>\u002Fdev\u002Fnull || echo \"See git log\")\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Focto-state.sh\" update_state \\\n  --history \"Files modified: $modified_files\"\n```\n\n\n## Terminal State\n\nThe Develop phase is complete ONLY when the implementation exists, the post-develop\ncheckpoint tag is created, and targeted tests pass fresh (see `skill-verification-gate`).\nThen invoke `flow-deliver` for validation. Do NOT declare the work done from here —\ncompletion claims belong to the Deliver phase after review.\n\n**Ready to build!** This skill is used after explicit invocation when users request implementation or building features.\n","> **Host: Codex CLI** — This skill was designed for Claude Code and adapted for Codex.\n> Cross-reference commands use installed skill names in Codex rather than `\u002Focto:*` slash commands.\n> Use the active Codex shell and subagent tools. Do not claim a provider, model, or host subagent is available until the current session exposes it.\n> For host tool equivalents, see `skills\u002Fblocks\u002Fcodex-host-adapter.md`.\n\n\n{{PREAMBLE}}\n\nLoad `skills\u002Fblocks\u002Fengineering-method-selection.md` from the installed plugin\nand apply only the methods relevant to this task. Preserve this entry point's\nexecution contract and output format. Read referenced skills as instructions;\ndo not invoke the current command recursively or add provider calls from a seat.\n\n## Portable feature boundary\n\nBefore implementation, select the existing feature and run its boundary adapter:\n\nUse the installation that loaded this skill for both execution and validation.\nIf `CLAUDE_PLUGIN_ROOT` is absent, replace every `OCTO_ROOT` fallback in this skill with\nthe absolute plugin root derived from this loaded skill's path. Do not infer it\nfrom the project directory or an existing stable link. The root helper refreshes\ninstalled copies while preserving another session's working root when called\nfrom an inferred development checkout.\n\n```bash\nOCTO_ROOT=\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\"\nbash \"$OCTO_ROOT\u002Fscripts\u002Fhelpers\u002Ffeature-workflow.sh\" boundary develop \"${OCTOPUS_FEATURE:-}\"\n```\n\nRead the bound project policy, relative source and digest. Ask the returned marker batch once through the available native host question tool. Keep skipped or noninteractive decisions open. Report the open count and defer only tasks whose current contract depends on a named unanswered decision. The runtime repeats these checks before actual task dispatch.\n\nTwo or more spec\u002Fplan\u002Ftasks artifacts trigger deterministic analysis automatically. A clean pass uses no extra model seat. Unresolved findings can trigger at most one independent bounded seat, shared across host and runtime for that artifact revision. Unavailable or failed semantic review warns and proceeds. Analysis reports exact source quotations and never rewrites artifacts.\n\nUse stable task IDs and metadata when present. The parent resolves paths, dependencies and current Git state at every wave. A hint cannot grant concurrency or permit overwriting user changes. Legacy inputs retain their existing planner and execution path.\n\n## Pre-Development: State Check\n\nBefore starting development:\n1. Read `.octo\u002FSTATE.md` to verify Define phase complete\n2. Update STATE.md:\n   - current_phase: 3\n   - phase_position: \"Development\"\n   - status: \"in_progress\"\n\n```bash\n# Verify Define phase is complete\nif [[ -f \".octo\u002FSTATE.md\" ]]; then\n  define_status=$(\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Focto-state.sh\" get_phase_status 2)\n  if [[ \"$define_status\" != \"complete\" ]]; then\n    echo \"⚠️ Warning: Define phase not marked complete. Consider running definition first.\"\n  fi\nfi\n\n# Update state for Development phase\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Focto-state.sh\" update_state \\\n  --phase 3 \\\n  --position \"Development\" \\\n  --status \"in_progress\"\n```\n\n\n## ⚠️ EXECUTION CONTRACT (MANDATORY - CANNOT SKIP)\n\nThis skill uses **ENFORCED execution mode**. You MUST follow this exact sequence.\n\n### STEP 1: Detect Work Context (MANDATORY)\n\nAnalyze the user's prompt and project to determine context:\n\n**Knowledge Context Indicators**:\n- Deliverable terms: \"PRD\", \"proposal\", \"presentation\", \"report\", \"strategy document\", \"business case\"\n- Business terms: \"market entry\", \"competitive analysis\", \"stakeholder\", \"executive summary\"\n\n**Dev Context Indicators**:\n- Technical terms: \"API\", \"endpoint\", \"function\", \"module\", \"service\", \"component\"\n- Action terms: \"implement\", \"code\", \"build\", \"create\", \"develop\" + technical noun\n\n**Also check**: Does project have `package.json`, `Cargo.toml`, etc.? (suggests Dev Context)\n\n**Capture context_type = \"Dev\" or \"Knowledge\"**\n\n#### Step 1b: Detect Dev Subtype (if Dev context)\n\nWhen context_type is Dev, determine the **subtype** to inject domain-appropriate quality guidance into the prompt sent to providers. Append the matching supplement text after the user's prompt.\n\n| Subtype | Trigger keywords | Quality supplement |\n|---------|-----------------|-------------------|\n| `frontend-ui` | \"page\", \"widget\", \"component\", \"UI\", \"HTML\", \"CSS\", \"form\", \"dashboard\", \"layout\" | See **frontend-ui enrichment** below. |\n| `cli-tool` | \"CLI\", \"command-line\", \"terminal\", \"script\", \"flag\", \"argument\" | Help text via --help flag. Meaningful exit codes (0 success, 1 user error, 2 system error). Stdin\u002Fstdout\u002Fstderr used correctly. Argument validation with clear error messages. |\n| `api-service` | \"API\", \"endpoint\", \"REST\", \"GraphQL\", \"gRPC\", \"server\", \"route\" | Input validation at boundaries. Consistent error response format. Auth\u002Fauthz on every endpoint. Rate limiting consideration. OpenAPI\u002Fschema documentation. |\n| `infra` | \"deploy\", \"terraform\", \"docker\", \"CI\", \"pipeline\", \"Kubernetes\", \"helm\" | Idempotent operations. Secrets never hardcoded. Rollback path documented. Health checks included. |\n| `data` | \"ETL\", \"pipeline\", \"migration\", \"schema\", \"database\", \"SQL\" | Idempotent migrations. Backup\u002Frollback strategy. Data validation at ingestion. |\n| `general` | Default if no subtype matches | No supplement — use base implementer persona only. |\n\n#### frontend-ui enrichment\n\nWhen `frontend-ui` subtype is detected, do TWO things:\n\n**A. Inject quality supplement into the prompt:**\nSelf-contained files preferred. Accessibility: ARIA labels, keyboard nav, 44px touch targets (WCAG 2.5.5). Safe DOM: createElement over innerHTML. Progressive enhancement: feature-detect APIs (navigator.share, localStorage) with fallbacks. Persist user prefs via localStorage.\n\n**B. Pull design intelligence from BM25 (if available):**\n\nBefore calling orchestrate.sh, check if the design intelligence engine exists and query it for relevant design context:\n\n```bash\nSEARCH_PY=\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fvendors\u002Fui-ux-pro-max-skill\u002Fsrc\u002Fui-ux-pro-max\u002Fscripts\u002Fsearch.py\"\nif [[ -f \"$SEARCH_PY\" ]]; then\n    # Detect relevant domains from the prompt\n    design_context=\"\"\n    # Style query — what visual style fits this task?\n    style_hit=$(python3 \"$SEARCH_PY\" \"\u003Cuser's task description>\" --domain style --top 1 2>\u002Fdev\u002Fnull || true)\n    [[ -n \"$style_hit\" ]] && design_context+=\"Design style suggestion: $style_hit\\n\"\n    # UX query — relevant UX patterns\n    ux_hit=$(python3 \"$SEARCH_PY\" \"\u003Cuser's task description>\" --domain ux --top 1 2>\u002Fdev\u002Fnull || true)\n    [[ -n \"$ux_hit\" ]] && design_context+=\"UX pattern: $ux_hit\\n\"\n    # Append to prompt if hits found\n    if [[ -n \"$design_context\" ]]; then\n        # Append design intelligence to the orchestrate.sh prompt\n        prompt=\"${prompt}\\n\\nDesign intelligence (from BM25 search):\\n${design_context}\"\n    fi\nfi\n```\n\nThis gives providers concrete design guidance (style direction, UX patterns) without requiring the user to run `\u002Focto:design-ui-ux` separately. If the search engine isn't installed, implementation proceeds with the quality supplement only.\n\n**How to apply:** When calling orchestrate.sh in Step 4, append the quality supplement (and design intelligence if available) to the prompt:\n```\norchestrate.sh develop \"\u003Cuser prompt>\\n\\nQuality requirements for this deliverable:\\n\u003Csupplement text>\\n\u003Cdesign intelligence if found>\"\n```\n\n**DO NOT PROCEED TO STEP 2 until context determined.** Context type (Dev vs Knowledge) and dev subtype determine which quality supplements and design intelligence to inject — wrong context wastes provider credits on irrelevant analysis.\n\n\n### STEP 2: Display Visual Indicators (MANDATORY - BLOCKING)\n\n**MANDATORY: You MUST use the native shell command tool to run this provider check BEFORE displaying the banner. Do NOT skip it. Do NOT assume availability.**\n\n```bash\nbash \"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fhelpers\u002Fcheck-providers.sh\"\n```\n\n**Use the ACTUAL results below. PROHIBITED: Showing only \"🔵 Claude: Available ✓\" without listing all providers.**\n\nIf `OCTO_ALLOWED_PROVIDERS` is set, treat it as the source of truth for which providers may participate. Providers filtered out by that allowlist are intentionally reported as unavailable; do not invoke or recommend them in the workflow.\n\n\n**Display this banner BEFORE orchestrate.sh execution:**\n\n**For Dev Context:**\n```\n🐙 **CLAUDE OCTOPUS ACTIVATED** - Multi-provider implementation mode\n🛠️ [Dev] Develop Phase: [Brief description of what you're building]\n\nProvider Availability:\n🔴 Codex CLI: ${codex_status} - Code generation and patterns\n🟡 Antigravity CLI: ${agy_status} - Alternative approaches\n🧭 Antigravity CLI: ${agy_status} - Additional external-model challenge\n🔵 Claude: Available ✓ - Integration and quality gates\n\n💰 Estimated Cost: 0.02-0.10 USD\n⏱️  Estimated Time: 3-7 minutes\n```\n\n**For Knowledge Context:**\n```\n🐙 **CLAUDE OCTOPUS ACTIVATED** - Multi-provider implementation mode\n🛠️ [Knowledge] Develop Phase: [Brief description of deliverable]\n\nProvider Availability:\n🔴 Codex CLI: ${codex_status} - Structure and framework application\n🟡 Antigravity CLI: ${agy_status} - Content and narrative development\n🧭 Antigravity CLI: ${agy_status} - Additional external-model challenge\n🔵 Claude: Available ✓ - Integration and quality review\n\n💰 Estimated Cost: 0.02-0.10 USD\n⏱️  Estimated Time: 3-7 minutes\n```\n\n**DO NOT PROCEED TO STEP 3 until banner displayed.** The banner shows users which providers will run and what costs they'll incur — starting API calls without this visibility violates cost transparency.\n\n\n### STEP 3: Read Prior State (MANDATORY - State Management)\n\n**Before executing the workflow, read any prior context:**\n\n```bash\n# Initialize state if needed\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" init_state\n\n# Set current workflow\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" set_current_workflow \"flow-develop\" \"develop\"\n\n# Get prior decisions (critical for implementation)\nprior_decisions=$(\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" get_decisions \"all\")\n\n# Get context from discover and define phases\ndiscover_context=$(\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" get_context \"discover\")\ndefine_context=$(\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" get_context \"define\")\n\n# Display what you found (if any)\nif [[ \"$discover_context\" != \"null\" ]]; then\n  echo \"📋 Discovery phase findings:\"\n  echo \"  $discover_context\"\nfi\n\nif [[ \"$define_context\" != \"null\" ]]; then\n  echo \"📋 Definition phase scope:\"\n  echo \"  $define_context\"\nfi\n\nif [[ \"$prior_decisions\" != \"[]\" && \"$prior_decisions\" != \"null\" ]]; then\n  echo \"📋 Implementing with decisions:\"\n  echo \"$prior_decisions\" | jq -r '.[] | \"  - \\(.decision) (\\(.phase)): \\(.rationale)\"'\nfi\n```\n\n**This provides critical context for implementation:**\n- Technology stack and patterns decided\n- Scope and requirements defined\n- Research findings to inform implementation\n- If **claude-mem** is installed, its MCP tools (`search`, `timeline`, `get_observations`) are available — use them to check for related past implementation patterns\n\n**DO NOT PROCEED TO STEP 4 until state read.**\n\n\n### STEP 4: Execute orchestrate.sh develop (MANDATORY - Use Bash Tool)\n\n**You MUST execute this command via the native shell command tool:**\n\n```bash\n# Keep execution and report validation in the same shell invocation.\nOCTO_ROOT=\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\"\nbash \"$OCTO_ROOT\u002Fscripts\u002Fhelpers\u002Fensure-plugin-root.sh\" || exit 1\nRUN_MARKER=$(mktemp) || exit 1\nRUN_STATUS=0\nOCTOPUS_WORKFLOW_INVOCATION_ID=\"${RUN_MARKER##*\u002F}\" \"$OCTO_ROOT\u002Fscripts\u002Forchestrate.sh\" develop \"\u003Cuser's implementation request>\" || RUN_STATUS=$?\nREPORT_STATUS=0\nSYNTHESIS_FILE=$(\"$OCTO_ROOT\u002Fscripts\u002Fhelpers\u002Fworkflow-result.sh\" develop \"$RUN_MARKER\" \"$RUN_STATUS\") || REPORT_STATUS=$?\nrm -f \"$RUN_MARKER\"\n[[ \"$REPORT_STATUS\" == 0 ]] || exit \"$REPORT_STATUS\"\necho \"VALIDATION PASSED: $SYNTHESIS_FILE\"\ncat \"$SYNTHESIS_FILE\"\n```\n\n**CRITICAL: You are PROHIBITED from:**\n- ❌ Implementing directly without calling orchestrate.sh — single-model implementation misses alternative approaches and edge cases that external providers surface through independent analysis\n- ❌ Writing code without multi-provider perspectives\n- ❌ Claiming you're \"simulating\" the workflow\n- ❌ Proceeding to Step 4 without running this command\n\n**You MUST use the native shell command tool to invoke orchestrate.sh.**\n\n#### What Users See During Execution (v7.16.0+)\n\nIf running in Claude Code v2.1.16+, users will see **real-time progress indicators** in the task spinner:\n\n**Phase 1 - External Provider Execution (Parallel):**\n- 🔴 Generating code and patterns (Codex)...\n- 🟡 Exploring alternative approaches (Antigravity)...\n\n**Phase 2 - Synthesis (Sequential):**\n- 🔵 Integrating and applying quality gates...\n\nThese spinner verb updates happen automatically - orchestrate.sh calls `update_task_progress()` before each agent execution. Users see exactly which provider is working and what it's doing.\n\n**If NOT running in Claude Code v2.1.16+:** Progress indicators are silently skipped, no errors shown.\n\n\n### STEP 5: Verify Execution (MANDATORY - Validation Gate)\n\n**After orchestrate.sh completes, verify it succeeded:**\n\nStep 4 checks the workflow exit status and validates a fresh\n`tangle-validation-*.md` report in this host's current session results directory.\nIt rejects missing reports and failed or unknown completion states. The validator\nhonors `CLAUDE_PLUGIN_DATA` and `CLAUDE_OCTOPUS_WORKSPACE`.\n\nUse the validated path printed by Step 4. If a later tool call starts a new shell,\nset `SYNTHESIS_FILE` to that exact path. Do not search for a different report\nor rerun the workflow to recover a shell variable.\n\n**If validation fails:**\n1. Report error to user\n2. Show logs from `~\u002F.claude-octopus\u002Flogs\u002F`\n3. DO NOT proceed with presenting results\n4. DO NOT substitute with direct implementation — fallback to single-model implementation skips the multi-provider synthesis that catches design flaws early\n\n\n### STEP 6: Update State (MANDATORY - Post-Execution)\n\n**After synthesis is verified, record implementation details in state:**\n\n```bash\n# Extract key implementation decisions from synthesis\nimplementation_approach=$(head -50 \"$SYNTHESIS_FILE\" | grep -A 3 \"## Implementation\\|## Approach\" | tail -3 | tr '\\n' ' ')\n\n# Record implementation decisions\ndecision_made=$(echo \"$implementation_approach\" | grep -o \"implemented\\|using [A-Za-z0-9 ]*\\|chose to\\|pattern:\" | head -1)\n\nif [[ -n \"$decision_made\" ]]; then\n  \"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" write_decision \\\n    \"develop\" \\\n    \"$decision_made\" \\\n    \"Multi-AI implementation consensus\"\nfi\n\n# Update develop phase context\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" update_context \\\n  \"develop\" \\\n  \"$implementation_approach\"\n\n# Update metrics\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Fstate-manager.sh\" update_metrics \"phases_completed\" \"1\"\n```\n\n**DO NOT PROCEED TO STEP 7 until state updated.**\n\n\n### STEP 7: Present Implementation Plan (Only After Steps 1-6 Complete)\n\nRead the synthesis file and present:\n- Recommended approach\n- Implementation steps\n- Code overview from all available perspectives\n- Quality gates results\n- Request user confirmation before implementing\n\n**After user confirms: Implement the solution using Write\u002FEdit tools**\n\n**Include attribution:**\n```\n*Multi-AI Implementation powered by Claude Octopus*\n*Providers: available external providers + 🔵 Claude*\n*Full implementation plan: $SYNTHESIS_FILE*\n```\n\n\n# Develop Workflow - Develop Phase 🛠️\n\n## ⚠️ MANDATORY: Context Detection & Visual Indicators\n\n**BEFORE executing ANY workflow actions, you MUST:**\n\n### Step 1: Detect Work Context\n\nAnalyze the user's prompt and project to determine context:\n\n**Knowledge Context Indicators** (in prompt):\n- Deliverable terms: \"PRD\", \"proposal\", \"presentation\", \"report\", \"strategy document\", \"business case\"\n- Business terms: \"market entry\", \"competitive analysis\", \"stakeholder\", \"executive summary\"\n\n**Dev Context Indicators** (in prompt):\n- Technical terms: \"API\", \"endpoint\", \"function\", \"module\", \"service\", \"component\"\n- Action terms: \"implement\", \"code\", \"build\", \"create\", \"develop\" + technical noun\n\n**Also check**: Does the project have `package.json`, `Cargo.toml`, etc.? (suggests Dev Context)\n\n**Step 1b: Detect Dev Subtype** — see EXECUTION CONTRACT Step 1b above for subtype table and quality supplements. Append the matching supplement to the prompt before calling orchestrate.sh.\n\n### Step 2: Output Context-Aware Banner\n\n**For Dev Context:**\n```\n🐙 **CLAUDE OCTOPUS ACTIVATED** - Multi-provider implementation mode\n🛠️ [Dev] Develop Phase: [Brief description of what you're building]\n📋 Session: ${CLAUDE_SESSION_ID}\n\nProviders:\n🔴 Codex CLI - Code generation and patterns\n🟡 Antigravity CLI - Alternative approaches\n🔵 Claude - Integration and quality gates\n```\n\n**For Knowledge Context:**\n```\n🐙 **CLAUDE OCTOPUS ACTIVATED** - Multi-provider implementation mode\n🛠️ [Knowledge] Develop Phase: [Brief description of deliverable]\n📋 Session: ${CLAUDE_SESSION_ID}\n\nProviders:\n🔴 Codex CLI - Structure and framework application\n🟡 Antigravity CLI - Content and narrative development\n🔵 Claude - Integration and quality review\n```\n\n{{VISUAL_INDICATORS}}\n\n\n**Part of Double Diamond: DEVELOP** (divergent thinking)\n\n```\n       DEVELOP (tangle)\n\n        \\         \u002F\n         \\   *   \u002F\n          \\ * * \u002F\n           \\   \u002F\n            \\ \u002F\n\n       Diverge with\n        solutions\n```\n\n## What This Workflow Does\n\nThe **develop** phase generates multiple implementation approaches using external CLI providers:\n\n1. **🔴 Codex CLI** - Implementation-focused, code generation, technical patterns\n2. **🟡 Antigravity CLI** - Alternative approaches, edge cases, best practices\n3. **🔵 Claude (You)** - Integration, refinement, and final implementation\n\nThis is the **divergent** phase for solutions - we explore different implementation paths before converging on the best approach.\n\n\n## When to Use Develop\n\nUse develop when you need:\n\n### Dev Context Examples\n- **Feature Implementation**: \"Build a user authentication system\"\n- **Code Generation**: \"Create an API endpoint for user registration\"\n- **Complex Builds**: \"Implement a caching layer with Redis\"\n- **Architecture Implementation**: \"Create a microservice for payment processing\"\n- **Integration Work**: \"Integrate Stripe payment processing\"\n\n### Knowledge Context Examples\n- **PRD Creation**: \"Build a PRD for the mobile onboarding feature\"\n- **Strategy Documents**: \"Create a market entry strategy for APAC\"\n- **Business Cases**: \"Build a business case for migrating to cloud\"\n- **Presentations**: \"Create an executive presentation on Q2 results\"\n- **Research Reports**: \"Build a competitive analysis report\"\n\n**Don't use develop for:**\n- Simple one-line code changes (use Edit tool)\n- Bug fixes (use debugging skills)\n- Code review tasks (use deliver-workflow or review skills)\n- Reading or exploring code (use Read tool)\n- Simple document edits (use Write tool)\n\n\n## Visual Indicators\n\nBefore execution, you'll see:\n\n```\n🐙 **CLAUDE OCTOPUS ACTIVATED** - Multi-provider implementation\n🛠️ Develop Phase: Building and developing solutions\n\nProviders:\n🔴 Codex CLI - Code generation and patterns\n🟡 Antigravity CLI - Alternative approaches\n🔵 Claude - Integration and refinement\n```\n\n\n## How It Works\n\n### Step 1: Invoke Tangle Phase\n\n```bash\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Forchestrate.sh\" develop \"\u003Cuser's implementation request>\"\n```\n\n### Step 2: Multi-Provider Implementation\n\nThe orchestrate.sh script will:\n1. Call **Codex CLI** with the implementation task\n2. Call **Antigravity CLI** with the implementation task\n3. You (Claude) contribute implementation analysis\n4. Synthesize approaches and recommend best path\n\n### Step 3: Review Quality Gates\n\nThe tangle phase includes automatic quality validation:\n- Code quality checks\n- Security scanning\n- Best practice validation\n- Implementation completeness\n\n### Step 3b: Handle Agent Continuation (v9.5 \u002F CC v2.1.77+)\n\nWhen orchestrate.sh output contains `AGENT_TEAMS_RESUME:${agent_id}:${task_id}:${role}:${phase}`:\n\n1. Send a message to the existing agent using `SendMessage({to: \"${agent_id}\", message: \"\u003Citeration prompt>\"})` — this auto-resumes a stopped agent automatically (CC v2.1.77+)\n2. Read the iteration prompt from the resume instruction JSON at `~\u002F.claude-octopus\u002Fagent-teams\u002F${task_id}.json`\n3. If SendMessage returns an error (agent GC'd or not found), fall back to creating a new agent with the full prompt from the instruction file\n4. After the agent responds, store its agentId via: `bridge_store_agent_id(task_id, agent_id)` in the bridge ledger\n\nNote: The `Agent(resume: ...)` parameter was removed in CC v2.1.77. Always use `SendMessage` for agent continuation.\n\n### Step 4: Read Results\n\nResults are saved to:\n```\n~\u002F.claude-octopus\u002Fresults\u002F${SESSION_ID}\u002Ftangle-validation-\u003Ctimestamp>.md\n```\n\n### Step 5: Implement Solution\n\nAfter reviewing all perspectives, implement the final solution using Write\u002FEdit tools.\n\n\n## Implementation Instructions\n\nWhen this skill is invoked, follow the EXECUTION CONTRACT above exactly. The contract includes:\n\n1. **Blocking Step 1**: Detect work context (Dev vs Knowledge)\n2. **Blocking Step 2**: Check providers, display visual indicators\n3. **Blocking Step 3**: Execute orchestrate.sh develop via native shell command tool\n4. **Blocking Step 4**: Verify synthesis file exists\n5. **Step 5**: Present implementation plan, get user confirmation\n6. **After confirmation**: Implement the solution using Write\u002FEdit tools\n\nEach step is **mandatory and blocking** - you cannot proceed to the next step until the current one completes successfully.\n\n### Task Management Integration\n\nCreate tasks to track execution progress:\n\n```javascript\n\u002F\u002F At start of skill execution\nTaskCreate({\n  subject: \"Execute develop workflow with multi-AI providers\",\n  description: \"Run orchestrate.sh develop for implementation\",\n  activeForm: \"Running multi-AI develop workflow\"\n})\n\n\u002F\u002F Mark in_progress when calling orchestrate.sh\nTaskUpdate({taskId: \"...\", status: \"in_progress\"})\n\n\u002F\u002F Mark completed ONLY after implementation finished\nTaskUpdate({taskId: \"...\", status: \"completed\"})\n```\n\n### Error Handling\n\nIf any step fails:\n- **Step 1 (Context)**: Default to Dev Context if ambiguous\n- **Step 2 (Providers)**: If all external providers are unavailable, suggest `\u002Focto:setup` and STOP\n- **Step 3 (orchestrate.sh)**: Show bash error, check logs, report to user\n- **Step 4 (Validation)**: If synthesis missing, show orchestrate.sh logs, DO NOT substitute with direct implementation\n\nNever fall back to direct implementation if orchestrate.sh execution fails. Report the failure and let the user decide how to proceed.\n\n### Implementation Plan Format\n\nAfter successful execution, present implementation plan with:\n   ```\n   # Implementation Plan: \u003Ctask>\n\n   ## Recommended Approach\n   [Your strategic recommendation based on all perspectives]\n\n   ## Implementation Steps\n   1. [Step-by-step implementation plan]\n   2. ...\n\n   ## Code Overview\n\n   ### Codex Approach\n   [Key implementation details from Codex]\n\n   ### Antigravity Approach\n   [Alternative considerations from Antigravity]\n\n   ### Final Implementation\n   [Your integrated solution]\n\n   ## Quality Gates\n   - [ ] Security validation: PASSED\u002FFAILED\n   - [ ] Best practices: PASSED\u002FFAILED\n   - [ ] Code quality: PASSED\u002FFAILED\n\n   Would you like me to proceed with implementation?\n   ```\n\n6. **Implement the solution**\n   After user confirms, use Write\u002FEdit tools to create the implementation:\n   - Create necessary files\n   - Write code based on synthesized approach\n   - Add tests if applicable\n   - Document the implementation\n\n\n## Example Usage\n\n### Example 1: Build Authentication System\n\n```\nUser: Build a user authentication system with JWT\n\nClaude:\n🐙 **CLAUDE OCTOPUS ACTIVATED** - Multi-provider implementation mode\n🛠️ Develop Phase: Building authentication system\n\n[Executes: \"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Forchestrate.sh\" develop \"Build a user authentication system with JWT\"]\n\n[After completion, reads synthesis and presents:]\n\n# Implementation Plan: User Authentication System\n\n## Recommended Approach\n\nBased on multi-provider analysis, I recommend a layered approach:\n1. Auth middleware for JWT validation\n2. Login\u002Fregister endpoints\n3. Token refresh mechanism\n4. Secure password hashing (bcrypt)\n\n## Implementation Steps\n\n1. Create `src\u002Fauth\u002Fmiddleware.ts` - JWT validation middleware\n2. Create `src\u002Fauth\u002Fcontroller.ts` - Login\u002Fregister endpoints\n3. Create `src\u002Fauth\u002Fservice.ts` - Business logic\n4. Create `src\u002Fauth\u002Futils.ts` - Token generation\u002Fvalidation\n5. Update `src\u002Froutes.ts` - Register auth routes\n\n## Code Overview\n\n### Codex Approach\n- Modern TypeScript with strict types\n- Express middleware pattern\n- Redis for token blacklisting\n- Comprehensive error handling\n\n### Antigravity Approach\n- Passport.js integration suggestion\n- Rate limiting on auth endpoints\n- Multi-factor auth consideration\n- Session management alternatives\n\n### Final Implementation\n- Hybrid: Modern TypeScript + Express patterns\n- JWT with refresh tokens\n- bcrypt password hashing\n- Rate limiting included\n- Optional MFA hooks for future\n\n## Quality Gates\n- ✅ Security validation: PASSED (bcrypt, secure tokens)\n- ✅ Best practices: PASSED (TypeScript, error handling)\n- ⚠️  Code quality: WARNING (consider adding request validation)\n\nWould you like me to proceed with implementation?\n\n[User: Yes, proceed]\n\n[Claude creates files and implements the solution]\n```\n\n### Example 2: Create API Endpoint\n\n```\nUser: Create an API endpoint for fetching user notifications\n\nClaude:\n🐙 **CLAUDE OCTOPUS ACTIVATED** - Multi-provider implementation mode\n🛠️ Develop Phase: Creating API endpoint\n\n[Executes tangle workflow]\n\n[Presents implementation plan with multi-provider perspectives]\n[Implements the endpoint after user confirmation]\n```\n\n\n## Quality Gates Integration\n\nThe tangle phase automatically runs quality checks via `.claude\u002Fhooks\u002Fquality-gate.sh`:\n\n```bash\n# Triggered after tangle execution (PostToolUse hook)\n.\u002Fhooks\u002Fquality-gate.sh\n```\n\n{{QUALITY_GATES}}\n\n\n## Integration with Other Workflows\n\nTangle is the **third phase** of the Double Diamond:\n\n```\nPROBE (Discover) → GRASP (Define) → TANGLE (Develop) → INK (Deliver)\n```\n\nAfter tangle completes, you may continue to:\n- **Ink**: Validate and deliver the implementation\n\nOr use standalone for implementation tasks.\n\n\n## Before Implementation Checklist\n\nBefore writing code, ensure:\n\n- [ ] All providers responded with implementation approaches\n- [ ] Quality gates evaluated (security, best practices, code quality)\n- [ ] User confirmed the implementation plan\n- [ ] File structure and architecture are clear\n- [ ] Dependencies identified and available\n- [ ] Tests planned (if applicable)\n\n\n## After Implementation: Auto Code Review & E2E Verification (MANDATORY)\n\n**After implementation completes and before presenting results to the user, you MUST launch two verification agents in parallel.** Do NOT skip this step or ask the user whether to run it — it is automatic.\n\n### Launch both agents simultaneously:\n\n**Agent 1 — Code Review (Sonnet):**\n```\nAgent(\n  model: \"sonnet\",\n  subagent_type: \"feature-dev:code-reviewer\",\n  background execution: true,\n  description: \"Code review: post-develop\",\n  prompt: \"Review the code changes from this development session. Focus on:\n1. Bugs, logic errors, security vulnerabilities\n2. Hidden dependencies or coupling issues\n3. Whether error handling covers failure modes\n4. Adherence to project conventions (check CLAUDE.md)\n\nCheck git diff for the changed files. Report only high-confidence issues.\"\n)\n```\n\n**Agent 2 — E2E Verification (Sonnet):**\n```\nAgent(\n  model: \"sonnet\",\n  background execution: true,\n  description: \"E2E test: post-develop\",\n  prompt: \"Run end-to-end verification of the development changes:\n1. Run the project's test suite (detect from package.json scripts, Makefile, or pyproject.toml)\n2. Verify no regressions in existing tests\n3. Check that new files are properly integrated (imported, registered, sourced)\n4. Verify the implementation matches the original task requirements\n\nReport: tests passed\u002Ffailed, any integration issues found.\"\n)\n```\n\n**After both agents complete:**\n- Present their findings to the user as part of the results\n- If the code reviewer found HIGH-confidence issues, flag them prominently\n- If tests failed, flag before the \"what next?\" prompt\n- Do NOT block on the review — present findings alongside results\n\nWHY: The user should never have to manually request a code review after development work. Fresh-eyes review from a different model (Sonnet vs Opus) catches issues the implementer is blind to. Running tests automatically catches regressions before the user discovers them.\n\n\n## After Implementation Checklist\n\nAfter writing code, ensure:\n\n- [ ] All files created\u002Fupdated\n- [ ] Code follows recommended patterns from synthesis\n- [ ] Code follows the project's existing commenting conventions (do not add comments unless asked or required by project style)\n- [ ] Security concerns addressed\n- [ ] Error handling implemented\n- [ ] Tests written (if applicable)\n- [ ] Lint\u002Ftypecheck commands run (detect from package.json, pyproject.toml, Cargo.toml, Makefile; if not found, ask the user)\n- [ ] If lint\u002Ftest commands were discovered and not documented in CLAUDE.md, suggest adding them\n- [ ] **Auto code review completed** (Sonnet agent)\n- [ ] **E2E verification completed** (Sonnet agent)\n- [ ] User notified of completion with review findings\n- [ ] Suggest running ink-workflow for validation\n\n\n## Cost Awareness\n\n**External API Usage:**\n- 🔴 Codex CLI uses your OPENAI_API_KEY (costs apply)\n- 🟡 Antigravity CLI uses your AGY_AUTH_TOKEN (costs apply)\n- 🔵 Claude analysis included with Claude Code\n\nTangle workflows typically cost 0.02-0.10 USD per task depending on complexity and code length.\n\n\n## Post-Development: Checkpoint\n\nAfter development completes:\n1. Run fresh targeted tests for the changed behavior through\n   `skill-verification-gate`; stop if they fail.\n2. Create checkpoint: `git tag octo-checkpoint-post-develop-$(date +%Y%m%d-%H%M%S)`\n3. Update `.octo\u002FSTATE.md` with completion and add a history entry with files modified.\n\n```bash\n# Enter this block only after fresh targeted tests pass.\ncheckpoint_tag=\"octo-checkpoint-post-develop-$(date +%Y%m%d-%H%M%S)\"\ngit tag \"$checkpoint_tag\" -m \"Post-develop checkpoint from embrace workflow\"\necho \"📌 Created checkpoint: $checkpoint_tag\"\n\n# Update state only after verification and checkpoint creation succeed.\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Focto-state.sh\" update_state \\\n  --status \"complete\" \\\n  --history \"Develop phase completed\"\n\n# Record files modified in this phase\nmodified_files=$(git diff --name-only HEAD~1 2>\u002Fdev\u002Fnull || echo \"See git log\")\n\"${CLAUDE_PLUGIN_ROOT:-${CODEX_PLUGIN_ROOT:-${HOME}\u002F.claude-octopus\u002Fplugin}}\u002Fscripts\u002Focto-state.sh\" update_state \\\n  --history \"Files modified: $modified_files\"\n```\n\n\n## Terminal State\n\nThe Develop phase is complete ONLY when the implementation exists, the post-develop\ncheckpoint tag is created, and targeted tests pass fresh (see `skill-verification-gate`).\nThen invoke `flow-deliver` for validation. Do NOT declare the work done from here —\ncompletion claims belong to the Deliver phase after review.\n\n**Ready to build!** This skill is used after explicit invocation when users request implementation or building features.\n",{"name":5,"description":7,"disable-model-invocation":49},true,"main","f620abaff09061f8c8b1f1559dca3fdd031a32e5","plugins\u002Fnyldn\u002Fclaude-octopus\u002Fskills\u002Fflow-develop",491,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins","https:\u002F\u002Fhol.org\u002Fplugins\u002Fbest-codex-plugins","https:\u002F\u002Fgithub.com\u002Fhashgraph-online","https:\u002F\u002Fhol.org","Open-source standards and developer tools for interoperable AI agents: discovery, identity, communication, payments, privacy and security.","2026-11-14 14:06:13",{"source":8,"skillName":5,"commands":61,"prompt":71,"zipAvailable":49},[62,65,68],{"tool":63,"command":64},"npx","npx skills add hashgraph-online\u002Fawesome-codex-plugins --skill flow-develop",{"tool":66,"command":67},"bunx","bunx skills add hashgraph-online\u002Fawesome-codex-plugins --skill flow-develop",{"tool":69,"command":70},"pnpm","pnpm dlx skills add hashgraph-online\u002Fawesome-codex-plugins --skill flow-develop","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fnyldn\u002Fclaude-octopus\u002Fskills\u002Fflow-develop\u002FSKILL.md",[73],{"path":74,"resourceType":75,"resourceTypeDesc":76,"sizeBytes":77,"sourceHtmlUrl":78},"agents\u002Fopenai.yaml","REFERENCE","参考文档",165,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fnyldn\u002Fclaude-octopus\u002Fskills\u002Fflow-develop\u002Fagents\u002Fopenai.yaml",[80,90,98,106,114,122],{"skillNo":81,"name":82,"nameZh":6,"description":83,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":84,"licenseFileUrl":85,"sourceHtmlUrl":86,"categories":87,"keywords":88,"publishedAt":89},"01M7DV332YQMGD8WTRJ9DZQXTB","worktable-review-with-annotations","Create, inspect, reply to, update, or resolve situated Worktable annotations. Use when feedback belongs to a Doc section, block, text range, or HTML Doc, or when the user asks to address existing review feedback and coordinate a requested artifact edit.","MIT","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-review-with-annotations\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-review-with-annotations\u002FSKILL.md",[],[],"2026-11-15 14:19:17",{"skillNo":91,"name":92,"nameZh":6,"description":93,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":84,"licenseFileUrl":94,"sourceHtmlUrl":95,"categories":96,"keywords":97,"publishedAt":89},"01M7DV2YGHCZVBEB1AHYB6CGHE","worktable-find-and-synthesize-context","Find, assess, and synthesize context from Worktable Docs and Records. Use when the user asks what a workspace contains, wants workspace-backed research, needs current facts reconciled across artifacts, or requests a cited synthesis without changing workspace state.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-find-and-synthesize-context\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-find-and-synthesize-context\u002FSKILL.md",[],[],{"skillNo":99,"name":100,"nameZh":6,"description":101,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":84,"licenseFileUrl":102,"sourceHtmlUrl":103,"categories":104,"keywords":105,"publishedAt":89},"01M7DV2TQN1H11ME2SAF9MXN7V","worktable-create-or-update-docs","Create, edit, rename, or maintain narrative Worktable Docs. Use for plans, decisions, briefs, notes, research syntheses, and other durable prose when the user requests a new artifact or a safe update to an existing one.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-create-or-update-docs\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-create-or-update-docs\u002FSKILL.md",[],[],{"skillNo":107,"name":108,"nameZh":6,"description":109,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":84,"licenseFileUrl":110,"sourceHtmlUrl":111,"categories":112,"keywords":113,"publishedAt":89},"01M7DV2NE452NEA8TXATKSNAJH","worktable-create-or-manage-records","Organize repeated structured information as durable Worktable Records. Use for research datasets, inventories, directories, pipelines, trackers, and other collections whose items need identity, fields, filtering, links, validation, or individual updates.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-create-or-manage-records\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-create-or-manage-records\u002FSKILL.md",[],[],{"skillNo":115,"name":116,"nameZh":6,"description":117,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":84,"licenseFileUrl":118,"sourceHtmlUrl":119,"categories":120,"keywords":121,"publishedAt":89},"01M7DV2HENRBR2KSCVM064WH0Y","worktable-create-or-manage-interactive-html","Create or improve durable visual and interactive artifacts in Worktable. Use for dashboards, trackers, calculators, mini apps, slide presentations, and other requests that benefit from a polished HTML Doc.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-create-or-manage-interactive-html\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-create-or-manage-interactive-html\u002FSKILL.md",[],[],{"skillNo":123,"name":124,"nameZh":6,"description":125,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":84,"licenseFileUrl":126,"sourceHtmlUrl":127,"categories":128,"keywords":129,"publishedAt":89},"01M7DV2CADR5R7RQFFYPNAGR9Y","worktable-collaborate-threads","Read, answer, and hand off messages through durable Worktable threads with users or other authenticated participants. Use when the user asks to check addressed messages, reply to a participant, continue a thread, request another agent's help, wait briefly for a response, or watch for comments and messages sent to the agent.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-collaborate-threads\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-collaborate-threads\u002FSKILL.md",[],[]]